Research Of Nonlinear Forecast On Minimum Ignition Energy And Minimum Ignition Temperature Of Mg-Al Alloy Dust

The active chemical properties of Mg-Al alloy easily leads to the occurrence of combustion and explosion, so it is a great threat to industrial safety. The minimum ignition temperature(MIT) and the minimum ignition energy(MIE), which are the characteristic parameters of dust explosion can measure the sensitivity of the dust explosion, and provide an important basis for the design of explosion technology and the selection of explosion equipment. This study aimed at providing the prediction models of MIT and MIE about Mg-Al alloy dust, and achieved the goal of getting their values by theoretical calculation but not test, Thus these prediction models can not only save our time and cost, but also be theoretical basis for enterprises to take timely measures to eliminate the cause of dust explosion potential unsafe factors.In this paper many groups of Mg-Al alloy dust were used to test MIT by Godbert-Greenwald furnace apparatus, many groups of Mg-Al alloy dust concentration were used to test MIE by related testing device. Based on the nonlinear fitting in regression analysis and evaluation of six indexes, the influence of concentration and particle size on MIT and MIE were fitted by R language and 1stOpt, and nonlinear prediction of MIT and MIE about Mg-Al alloy dust were achieved. This paper studied the correlation between MIT and MIE were confirmed, and the model of them was got by regression analysis.The fitting results were as follows:(1)The fitting model for MIT was got. In this model, the coefficient of determination(R2) was 96.93%,higher than 95%; the total relative error(TRE) was 0.17%, and the mean systematic error(MSE) was 0.12%,which were less than ±3% of the modeling requirements.(2)The fitting model for MIT was got. In this model, the coefficient of determination(R2) was 94.93%,higher than 95%; the total relative error(TRE) was- 0.12%, and the mean systematic error(MSE) was-0.79%,which were less than ±3% of the modeling requirements.(3)The fitting model for the relationship of MIT and MIE was got. In this model, the coefficient of determination(R2) was 98.50%,higher than 95%; the total relative error(TRE) was0.09%, and the mean systematic error(MSE) was 0.02%,which were less than ±3% of the modeling requirements.